Arsenic trioxide (ATO) induces apoptosis and inhibits cell growth and angiogenesis via mechanisms that are not yet fully understood. ATO is available as an intravenously injectable formulation (Trisenox) and is approved for the treatment of acute promyelocytic leukemia [10]. Arsenic trioxide-coated stents have been shown to inhibit neointimal proliferation in preclinical models of in-stent stenosis [11, 12, 13].

Glucocorticoids have an inhibitory effect on inflammatory cell activation, cytokine release, platelet adhesion, smooth muscle cell proliferation, and collagen synthesis [14, 15]. In view of these properties, it is tempting to speculate that glucocorticoids might inhibit restenosis after PCIs. Oral glucocorticoid therapy was shown to reduce the occurrence of restenosis after angioplasty [16]. Both preclinical and clinical studies investigating local intravascular administration of glucocorticoids via coated stents found beneficial effects with respect to restenosis inhibition and target lumen revascularization (TLR) [14, 17].

Bortezomib is a selective proteasome inhibitor that is approved for the treatment of multiple myeloma. It inhibits proliferation and induces apoptosis of metabolically active, rapidly dividing cells at very low concentrations [18]. Systemically and locally delivered bortezomib reduced neointimal proliferation in preclinical models of arterial injury [19, 20].

Green tea extract contains a number of catechins, including epigallocatechin-3-gallate (EGCG), epigallocatechin (EGC), epicatechin-gallate (ECG), and epicatechin (EC). EGCG was shown to have potential beneficial effects on vascular function via vasodilatory and antiplatelet activity [21, 22, 23, 24]. Preclinical studies in rats indicate that EGCG reduces intimal hyperplasia in models of intravascular injury [25, 26].

Similar to taxol, the epothilones (A and B) are microtubule-stabilizing agents causing cell cycle arrest [27, 28]. Locally administered epothilones suppress neointimal thickening in the rat balloon injury model [29, 30, 31]. In our experiments we used fantolon (Bayer-Schering Pharma AG, C${}_{30}$ H${}_{41}$ N O${}_6$ S) because of its known efficacy at very low dose/concentration.

Methotrexate (MTX), an analogue of the B vitamin folic acid, is a first-line synthetic disease-modifying antirheumatic drug for patients with rheumatoid arthritis (RA) and other autoimmune diseases [32]. Systemic (IV) treatment with methotrexate showed beneficial effects in a rabbit model of in-stent neoatherosclerosis [33]. Studies in RA patients at high risk for cardiovascular diseases suggest that MTX has potential protective effects. It might exert beneficial effects on endothelial function, regulating immune responses and inflammation pathways responsible for the development of atherosclerosis and thrombosis [34, 35], possibly diminishing in-stent restenosis after PCIs [36].

Thalidomide inhibits neoangiogenesis, which contributes to intravascular plaque formation. It is a partial inhibitor of the biosynthesis of the key proinflammatory cytokine tumor necrosis factor alpha (TNF$\alpha$) [37, 38]. TNF$\alpha$ is produced by a number of cells, including macrophages, neutrophils, endothelial cells, and vascular smooth muscle cells (VSMCs) and was shown to play a pivotal role in restenosis [39]. In preclinical models, systemic and local delivery of thalidomide reduced injury-induced neointima formation [40, 41].

First, we tested the solubility of the selected drugs and their stability in various organic solvents at room temperature. Formulations containing a mixture of the active substance, solvents and, where appropriate, an excipient (Table 1) were coated onto percutaneous transluminal coronary angioplasty (PTCA) balloon catheters (3.5 mm $\times$ 20 mm) by dipping the balloons into the coating solution or by dispersing a precisely defined volume on the balloon surface using a Hamilton syringe (Fig. 1). After complete drying for 24 h at room temperature, coated balloons were visually inspected for appearance, homogeneity, and adherence of the drug coating. The amount of drug on the balloon was analyzed, in most cases by high-pressure liquid chromatography (HPLC) (Table 1).

Fig. 1.

Drug coated, inflated PCTA balloon.

All in vivo experiments were performed at the IMTR (Institute of Medical Technology and Research, Saxony-Anhalt) and conducted in accordance with the guidelines for animal experiments and approved by the animal protection committee of Saxony-Anhalt, Germany. Balloon catheters used for in vivo studies were sterilized with ethylene oxide according to a validated protocol. Studies in the porcine coronary in-stent stenosis model were conducted as described in Kelsch et al. [42]. Briefly, after sedation, anesthesia and analgesic treatment, the common carotid artery was surgically exposed and an intra-arterial sheath was introduced. The coronary arteries were visualized with a Siemens AXIOM Artis zee fluoroscope using a 6F Judkins L3.5 catheter and Ultravist®-370 as a contrast agent. Coronary angiograms were recorded after intracoronary nitroglycerin administration. After coronary imaging, a suitable segment of the left coronary artery circumflex ramus (CX), left anterior descending artery (LAD), and right coronary artery (RCA) was treated in each pig.

Drug transfer from the balloon surface to the vessel wall was examined as follows: coronary artery segments were labeled with a marker stent. Areas proximal to the implanted marker stent were treated with the coated balloons applying approximately 20% overstretch to ensure vessel wall contact. Balloons were left expanded for 1 minute, deflated, and retracted. The pigs were euthanized immediately after the last treatment in deep anesthesia using supersaturated potassium chloride. The treated vessel segments including the marker stents were dissected for drug analysis.

Neointimal proliferation was induced by implantation of balloon-expandable bare metal stents using either stents mounted onto balloons (coated or uncoated control) with a hand crimper HH100–101 (Machine Solutions, Flagstaff, AZ) or commercially available premounted bare metal stents on PTCA balloon catheters before postdilatation with experimental DCBs. Stents were expanded in coronary arteries applying approximately 20% overstretch to the vascular segment. Blood flow and lumen dimensions were monitored before and after stent implantation by coronary angiography (see below). Four weeks after the intervention, follow-up angiography of stented segments was done. Then, animals were euthanized in deep anesthesia, and the treated vessel segments were dissected and fixed with 4% buffered formalin for histomorphometric analysis.

Samples containing extraction solvent and unused balloons or balloons after use were intensely shaken on a vortexer for at least 30 seconds, followed by treatment in an ultrasound bath for 30 minutes and centrifugation for 10 minutes. Except for arsenic trioxide, drug concentrations were determined by HPLC with UV detection (Table 2). Drug analysis by HPLC was done on reversed-phase C18 columns (5 $\mu$m, 4.6 mm $\times$ 250 mm) or Core-shell columns (2.5 $\mu$m, 4.6 mm $\times$ 150 mm). Mobile phases were selected according to the drug. Standards were run together with the samples. For drug extraction from tissues, a defined volume of ethanol was added to achieve an ethanol concentration of $>$50%. Arsenic trioxide was extracted with aqueous NH${}_4$OH. The samples were homogenized (Precelly 24 Dual Homogenizer, PEQLAB Biotechnologie GmbH, Erlangen, Germany) and extracted by 30 minutes ultrasound treatment at room temperature and then centrifuged for 10 minutes at 17,500 $\times$ g. Except for arsenic trioxide, drug concentrations were determined in the supernatant by HPLC with UV detection.

Quantitative coronary angiography (QCA) was done before, during, and after balloon inflation and stent implantation. Four weeks after stent implantation, follow-up QCA was performed at the respective implant sites to assess the primary endpoint late lumen loss (LLL, defined as minimal lumen diameter at implantation (MLD) – minimal lumen diameter at 4-weeks follow-up (MLD FU)) and in-stent stenosis % (defined as (1 – MLD FU / reference diameter (RFD)) $\times$ 100) of stented segments. Angiograms were assessed by two experienced observers blinded to the treatment groups using the CAAS II Research System (Pie Medical, The Netherlands) (Fig. 2).

Fig. 2.

Angiograms showing a coronary vessel pre- and post treatment with a DCB and at 4-weeks follow-up (FU).

The target segments of treated coronary arteries were embedded in methyl-methacrylate, cut with a coping saw, polished, and glued on acrylic plastic slides. Cross-sections from proximal, mid and distal portions of the stent were stained with hematoxylin and eosin (H&E) and/or Masson-Goldner and digitalized. From these images, histomorphometric measurements of lumen and vessel wall area (defined as vessel area – lumen area) were recorded. Injury scores were assigned as described by Schwartz et al. [44], and inflammation was graded individually for each stent strut using the score introduced by Kornowski et al. [45]. The injury score and the inflammatory score for each cross-section were calculated as the sum of the individual injury and inflammatory scores divided by the number of struts in the examined section.

QCA-derived parameters were compared across treatment groups. Histomorphometric variables of the three cross-sectional planes were averaged to obtain a mean value per stent. Continuous variables were compared by Kruskal-Wallis one-way ANOVA test using Graphpad Prism (Version 6, GraphPad Software, San Diego, CA, USA). Data are presented as mean $\pm$ SD.

Compounds were completely dissolved in the aforementioned mixtures and coated on plain 3.5 mm $\times$ 20 mm balloons. The results of drug content analysis are compiled in Table 2, given as $\mu$g/mm${}^2$ of the total balloon surface (approx. 220 mm${}^2$). Coating composition (solvents and excipients) were selected to accomplish transfer of $\ge$5% of the dose on the balloon, which corresponds to the lowest percentage of PTX transferred from clinically proven DCBs [42, 46].

On the basis of the in vitro and in vivo studies, the following drugs were selected to be tested with respect to their efficacy in reducing in-stent stenosis in the porcine coronary overstretch and stent implantation model: arsenic trioxide, betamethasone dipropionate, bortezomib, epigallocatechin-gallate, fantolon, methotrexate, and thalidomide. Both uncoated and PTX-coated balloons, for which inhibition of restenosis is known from controlled clinical trials [47, 48], served as controls. Outcome of balloon treatment in the porcine in-stent stenosis was assessed from in vivo coronary angiography and ex vivo histological analysis. First, pooled QCA data at 4-week follow-up for uncoated balloons from multiple independent studies were compared with data for the three different PTX coating formulations, namely Paccocath, InPact, and experimental amorphous PTX (see Table 2). All three PTX formulation variants significantly reduced angiographic in-stent LLL compared to the pooled uncoated balloon group (Fig. 3). LLL reduction was not significantly different among the three PTX variants.

Fig. 3.

QCA in-stent late lumen loss (LLL) at 4-week follow-up. Balloons coated with different PTX formulations significantly reduce LLL compared to uncoated balloons. Data are shown as mean $\pm$ SD and data points, ****p $\le$ 0.0001, **p $\le$ 0.01 vs. uncoated control, Kruskal-Wallis test with Dunn’s post hoc comparison.

To elucidate a potential inhibition of neointimal proliferation of the different novel balloon coatings, we compared data points from multiple experiments investigating the drug candidates with the pooled results achieved with uncoated control balloons and PTX formulation variants. For all studies, quantitative coronary angiography revealed similar angiographic baseline parameters in each treatment group within each study versus the uncoated control. Except for treatment with PTX-coated balloons, angiographic parameters at 4 weeks after implantation did not show a significant reduction of in-stent stenosis in any of the tested drug-coated balloon groups compared to uncoated controls (Table 3, Fig. 4). Unlike treatment with PTX-coated balloons, none of the new coatings we investigated inhibited neointimal proliferation, i.e., QCA showed no decrease in LLL or in-stent stenosis or increase in minimal lumen diameter compared to treatment with uncoated balloons.

Fig. 4.

QCA in-stent late lumen loss (LLL) at 4-week follow-up. Following treatment with PTX-coated balloons, LLL is significantly lower compared with the pooled results achieved with uncoated balloons. For all other coatings tested, results are not significantly different from those of the uncoated control group. Data are shown as mean $\pm$ SD and data points, ****p $\le$ 0.0001 vs. uncoated control, Kruskal-Wallis test with Dunn’s post hoc comparison.

Histological analysis of stented cross-sections showed no reduction of neointimal area for any balloon coatings except PTX compared to the uncoated control group (Table 4, Fig. 5). These results confirm the ineffectiveness of the tested balloon coatings with respect to inhibition of neointimal proliferation.

Fig. 5.

Histomorphometric in-stent vessel wall area at 4-week follow-up. Following treatment with PTX-coated balloons, the vessel wall area (including neointima) is significantly smaller than after treatment with uncoated balloons. For all other coatings tested, results are not significantly different from those of the uncoated control group. Data are shown as mean $\pm$ SD and data points, PTX mean – dotted line, ****p $\le$ 0.0001 vs. uncoated control, Kruskal-Wallis test with Dunn’s post hoc comparison.

Cross-sections of stented artery segments showed significant neointimal thickening and lumen narrowing for all balloon coatings investigated except for the PTX group. Nearly all sections from arteries treated with PTX-coated balloons exhibited only a thin cellular layer covering the stent struts (Fig. 6).

Fig. 6.

Representative H&E-stained histological cross-sections of stented artery segments at 4-week follow-up. BDP, betamethasone dipropionate.

None of the tested devices, coatings, or applications caused any signs of obvious toxicity during the experiments. None of the drugs had to be excluded from further testing because of recognizable local or systemic toxicity.